Open Access Article
Emma G. L.
Robert
and
Jerome
Waser
*
Laboratory of Catalysis and Organic Synthesis, Institute of Chemical Sciences and Engineering, Ecole Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland. E-mail: jerome.waser@epfl.ch
First published on 23rd May 2025
We report the diastereoselective and enantioselective synthesis of thio-substituted cyclobutanes via a sulfa-Michael addition using cyclobutenes. In the presence of DBU, various thio-cyclobutane esters and amides were obtained in up to quantitative yield and >95
:
5 dr. Using a chiral chinchona-based squaramide bifunctional acid–base catalyst and an N-acyl-oxazolidinone-substituted cyclobutene, thio-cyclobutanes were obtained with high yield and enantioselectivity (er up to 99.7
:
0.3).
![]() | ||
| Scheme 1 S-Substituted cyclobutanes in medicinal chemistry (A), synthetic strategies (B), and this work (C). | ||
There are currently four main strategies for accessing sulfur-substituted cyclobutanes: cyclization of acyclic precursors by deprotonation in the α-position to a sulfone, sulfoximine or sulfoxide and the subsequent substitution reaction (Scheme 1 and eqn (1));7 thermal-catalyzed,8 Lewis acid-catalyzed9 or photo-catalyzed10 [2 + 2] cycloaddition (Scheme 1 and eqn (2)); incorporating a sulfur group into an existing cyclobutene, either via radical intermediates (Scheme 1 and eqn (3))11–14 or through Michael addition of a thio nucleophile (Scheme 1 and eqn (4)).5b,6,8,15,16 The ring expansion of cyclopropanes to sulfur-substituted cyclobutanes was also explored, but it occurred with low efficiency.17
Concerning the radical sulfur addition pathway, Witiak and co-workers reported in 1972 the first two examples of Giese addition of thioacetic acid and benzyl thiol radicals generated by light irradiation of cyclobut-1-ene-1-carboxylic acid.11 These conditions led to a mixture of isomers in the case of benzyl thiol. In 2020, Gouverneur and co-workers used sulfonyl12 and sulfamoyl13 chlorides for the hydrosulfonylation and hydrosulfamoylation of activated alkenes under visible light catalysis. They described five examples using cyclobutene esters and amides as alkenes to obtain sulfoxide-substituted cyclobutanes majorly with a cis-configuration. In the same year, Landais and co-workers reported the sulfonylcyanation of chiral cyclobutenes under light irradiation.14 In this case, a trans-relationship between the N-acyl-oxazolidinone and the sulfonyl group was obtained.
Surprisingly, although the Michael addition of nitrogen nucleophiles to cyclobutenes has been exploited,18 its application to sulfur-based nucleophiles has remained largely unexplored. Ciabatti and co-workers reported that the addition of thioacetic acid to cyclobut-1-ene-1-carboxylic acid was also possible without generation of the radical but proceeded with only 7
:
3 dr.5b Probably, due to the low selectivity, this method found only occasional use for the generation of simple building blocks for medicinal chemistry projects.6 Besides addition to cyclobut-1-ene-1-carboxylic acid, the only other cases of Michael addition involve a non-selective addition of glutathione to a conjugated ketone in the context of metabolic intermediate studies,15a as well as additions to cyclobutene esters on a biased bicyclic system15b and on a polyhalogenated derivative.8 In 2019, Aitken and co-workers reported access to β-sulfinyl cyclobutane amides via a rearrangement of α-sulfinyl precursors.16 The scope was limited to aryl sulphoxides, and the products were obtained in a diastereomeric mixture due to the stereogenic sulphoxide. Overall, highly diastereoselective Michael additions of thio-nucleophiles to cyclobutene esters remain extremely rare, and no enantioselective method has been reported yet. This is surprising considering that numerous examples of enantioselective sulfa-Michael additions using less strained substrates have been reported over the past two decades, primarily relying on the use of chiral metal complexes or organocatalysts.19
Herein, we describe the first highly diastereoselective (>95
:
5 dr) and enantioselective (up to 99.7
:
0.3 er) synthesis of thio-cyclobutanes. Starting from commercially available thiols and readily available cyclobutenes, our method enables the efficient formation of a wide range of thio-substituted cyclobutane esters and amides. High diastereoselectivities were first achieved just by using DBU as a base. With a chiral chinchona squaramide catalyst, thio-cyclobutanes were then accessed with high enantioselectivity. This novel method therefore offers for the first time control over both diastereoselectivity and enantioselectivity for the synthesis of 1,2-substituted thio-cyclobutane esters, offering efficient access to a currently underexploited chemical space for applications in medicinal chemistry.
:
50 dr). Switching K2CO3 for NEt3 (entry 2) or TMG (entry 3) increased the yield but resulted in a low dr (95% yield, 51
:
49 dr and 90% yield, 64
:
36 dr, respectively). TBD offered the desired product in 69% yield and 72
:
28 dr (entry 4). In contrast, DBU provided the trans-product in quantitative yield and >95
:
5 dr (entry 5). Changing the solvent from MeCN to EtOAc resulted in the same yield but lower dr (82
:
18) (entry 6). Lowering the reaction time to 1 or 2 hours did not affect the yield but lowered the diastereoselectivity to 64
:
36 after 1 hour (entry 7) and to 71
:
29 (entry 8) after 2 hours, suggesting thermodynamic control of the dr. Finally, conducting the reaction under an air atmosphere (entry 9) or with HPLC-grade MeCN (entry 10) resulted in a lower diastereoselectivity (86
:
14 dr and 89
:
11 dr, respectively), probably due to the presence of water, which can slow down the epimerization.
| Entry | Base | Solvent | Time | Yieldb | drc |
|---|---|---|---|---|---|
| a 1.0 equiv. thiol 1a (0.1 mmol), 1.1 equiv. cyclobutene 2a, 1.1 equiv. base, rt. b 1H NMR of the crude mixture with dibromomethane as an internal standard. c Measured from the crude 1H NMR. | |||||
| 1 | K2CO3 | MeCN | 18 h | 72% | 50 : 50 |
| 2 | NEt3 | MeCN | 18 h | 95% | 51 : 49 |
| 3 | TMG | MeCN | 18 h | 90% | 64 : 36 |
| 4 | TBD | MeCN | 18 h | 69% | 72 : 28 |
| 5 | DBU | MeCN | 18 h | Quant. |
>95 : 5
|
| 6 | DBU | EtOAc | 18 h | Quant. | 82 : 18 |
| 7 | DBU | MeCN | 1 h | Quant. | 64 : 36 |
| 8 | DBU | MeCN | 2 h | 98% | 71 : 29 |
| 9 | DBU | MeCN + air | 18 h | 81% | 86 : 14 |
| 10 | DBU | HPLC-MeCN | 18 h | 97% | 89 : 11 |
:
5 dr on a 0.3 mmol scale. Scaling up the reaction to 1 mmol offered the desired product in 81% yield and an identical dr. A 4-fluorine or 3-chlorine substituent on the benzene moiety was well tolerated and led to 3b (88% yield and >95
:
5 dr) and 3c (82% yield and >95
:
5 dr), respectively. Other electron-withdrawing groups could be introduced, such as a CF3 (3d; 89% yield and >95
:
5 dr), a NO2 (3e; 58% yield and >95
:
5 dr) or a methyl ester (3f; 70% yield and >95
:
5 dr) group. For the latter, 3.0 equivalents of DBU and heating to 80 °C were necessary to obtain the desired product with high diastereoselectivity (method (B)). These conditions showed improved diastereoselectivity for all the thiols not bearing an electron-withdrawing group. Simple thiophenol provided 3g in 84% yield and >95
:
5 dr. tert-Butylthiophenol led to 3h in 90% yield and >95
:
5 dr, while 2,6-dimethylthiophenol offered 3i in 93% yield and >95
:
5 dr. A chromenone-substituted thiol was tolerated, and 3j was obtained in 93% yield and 88
:
12 dr with method (A) and in 33% yield and >95
:
5 dr with method (B). Methoxy-substituted thiophenol provided 3k in quantitative yield and 89
:
11 dr. A free alcohol, an acetamide and a free amine led to 3l (68% yield and >95
:
5 dr), 3m (88% yield and >95
:
5 dr) and 3n (88% yield and 94
:
6 dr), respectively. The X-ray structure of 3'l confirmed the molecular structure and relative configuration of the minor cis-diastereoisomer.21
Heteroaromatic-substituted thiols could also be introduced; 2-pyridine as well as furan-substituted thiol provided 3o (70% yield and >95
:
5 dr) and 3p (76% yield and >95
:
5 dr), respectively. Moreover, benzyl thiol provided 3q in quantitative yield and 93
:
7 dr. Similarly, methoxy- and chloro-substituted benzyl thiols could be introduced, giving 3r (96% yield and 92
:
8 dr) and 3s (89% yield and >95
:
5 dr), respectively. Furan-2-ylmethanethiol provided 3t (80% yield and 94
:
6 dr). A primary alkyl-substituted thiol led to 3u (quant. yield and >95
:
5 dr), while a secondary thiol gave 3v (quant. yield, 93
:
7 dr). The sterically hindered adamantyl thiol led to 3w in quantitative yield and 29
:
71 dr with method (A) and 94% yield and 92
:
8 dr with method (B). Thioacetic acid provided 3x in 71% yield and 67
:
33 dr with method (A). In this case, the use of method (B) could not improve the diastereoselectivity as decomposition was observed. Finally, replacement of the benzyl ester was implemented with an n-butyl ester (3y; 96% yield and 94
:
6 dr), a Weinreb amide (3z; 93% yield and >95
:
5 dr) or a benzocaine-substituted amide (3aa; 68% yield and >95
:
5 dr).
![]() | ||
| Fig. 1 Structures of chiral bifunctional organocatalysts used in the optimization of the enantioselective sulfa-Michael addition. | ||
Using 2 mol% of the commercially available thiourea cinchona catalyst Cat*1 in toluene at room temperature overnight, we were pleased to obtain the desired product in 92% yield, modest diastereoselectivity of 79
:
21 and good enantiomeric ratio of 96
:
4 for the major diastereoisomer (Scheme 3 and Table 2, entry 1), while the er of the minor diastereoisomer was lower (59
:
41 er). To improve the dr, 3.0 equivalents of DBU were added and the reaction mixture was heated to 70 °C for one hour. An improvement of dr was observed (94
:
6 dr compared to 79
:
21 dr), but the er dropped to 57
:
43, suggesting that the epimerization occurred via a retro-Michael/Michael-addition process rather than a simple deprotonation. This hypothesis was confirmed by adding a second thiol to the crude reaction mixture after full conversion, resulting in a mixture of the two thiol-substituted products (see Section 4.1 in the ESI†).
| Entry | Catalyst | Yieldb | drc | erd |
|---|---|---|---|---|
| a 1.0 equiv. thiol 1a (0.1 mmol), 1.2 equiv. cyclobutene 2e, 2 mol% Cat*, toluene [0.1 M], rt, 18 h. b 1H NMR of the crude mixture with dibromomethane as an internal standard. c Measured from the crude 1H NMR. d Measured from the SFC chromatogram. | ||||
| 1 | Cat*1 | 92% | 79 : 21 |
96 : 4 |
| 2 | Cat*2 | Quant. | 85 : 15 |
96 : 4 |
| 3 | Cat*3 | Quant. | 71 : 29 |
6 : 94 |
| 4 | Cat*4 | 86% | 48 : 52 |
5 : 95 |
| 5 | Cat*5 | 98% | 82 : 18 |
92 : 8 |
| 6 | Cat*6 | Quant. | 55 : 45 |
38 : 62 |
| 7 | Cat*7 | 84% | 66 : 34 |
89 : 11 |
| 8 | Cat*8 | 99% |
89 : 11
|
98 : 2
|
| 9 | Cat*9 | 98% | 91 : 9 |
90 : 10 |
We then turned our attention to other chiral organocatalysts in order to further improve the diastereoselectivity under kinetic control. Cat*2 with a hydrogenated quinuclidine substituent led to an improvement in dr (85
:
15) and the same er (Table 2, entry 2). Cat*3 with an inverted urea stereocenter gave a quantitative yield, but lower dr (71
:
29) and er (6
:
94) than Cat*1. Removing the methoxy group of the quinoline backbone (Cat*4) led to a low dr (entry 4). Takemoto's thiourea (Cat*5–6) and urea (Cat*7) type of catalysts led to a decrease in diastereoselectivity and enantioselectivity (entries 5–7). Squaramide-based cinchona Cat*8 gave the best results with 99% yield, 89
:
11 dr and 98
:
2 er (entry 8). Replacement of the chinchona by a chiral diamine moiety (Cat*9) gave a similar yield (98%), a higher dr (91
:
9), but a lower enantioselectivity (90
:
10 er) (entry 9). Cat*8 was then selected as the best catalyst, and further fine-tuning of the conditions was performed, but no significant improvement of neither the dr nor the er could be obtained (see the ESI†). In addition, performing the reaction with ester-substituted cyclobutene 2a led to the formation of thiocyclobutane 3a in a racemic form, confirming the essential role of the oxazolidinone auxiliary.
:
9 dr, and 98
:
2 er). Among aromatic thiols, an ester electron-withdrawing group, a neutral bulky tert-butyl group, and a methoxy electron-donating group all smoothly underwent the transformation with high diastereo- and enantioselectivity; 4b (93% yield, >95
:
5 dr, and 97
:
3 er), 4c (82% yield, 91
:
9 dr, and 99.7
:
0.3 er), and 4d (quant. yield, 88
:
12 dr, and 99
:
1 er) were obtained, respectively. The X-ray structure of 4b confirmed the molecular structure and absolute configuration of the major trans-enantiomer as (S-,R-), in accordance with Houk's Brønsted acid–hydrogen bonding stereoinduction model (see the ESI, Fig. S1†).22 2-Pyridine-substituted thiol produced 4e (95% yield, 61
:
39 dr, and 96
:
4 er). Aliphatic thiols were also suitable in the enantioselective transformation. Benzyl thiol 4f (quant. yield, 68
:
32 dr, and 98
:
2 er) and substituted benzyl thiol 4g (62% yield, 91
:
9 dr, and 97
:
3 er) and 4h (quant. yield, >95
:
5 dr, and 98
:
2 er) were tolerated. Furan-2-ylmethanethiol produced 4i (79% yield, 90
:
10 dr, and 99
:
1 er), while 2-phenylethane-1-thiol gave 4j (92% yield, 85
:
15 dr, and 95
:
5 er). However, bulkier thiols such as cyclohexyl thiol and adamantyl thiol were not compatible with the reaction and provided the desired product in low yields (14% and <5%, respectively).
Moreover, other ester derivatives were synthesized from enantioenriched 4a, including methyl ester 6 (95% yield and 97
:
3 er) (eqn (3)) and the tert-butyl ester 7, obtained via saponification followed by esterification in 42% yield and 97
:
3 er (eqn (4)). Enantioenriched alcohol 8 can be obtained in 92% yield and with unchanged er by reduction of the oxazolidinone auxiliary using NaBH4 (eqn (5)). Furthermore, the Weinreb amide analog 3z was obtained by reaction with the amine chloride salt in the presence of base and a catalytic amount of Yb(OTf)3 (eqn (6)) and was subsequently converted into the corresponding methyl ketone 9 in 91% yield without loss of er (eqn (7)).
Footnote |
| † Electronic supplementary information (ESI) available: Optimization details, experimental procedures, characterization data and NMR spectra of new compounds. Raw data for NMR, IR, and HRMS will be made freely available on the platform Zenodo. CCDC 2383043 and 2415842. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d5sc01727k |
| This journal is © The Royal Society of Chemistry 2025 |